MEMORANDUM

No 14/2006

Hilde Bojer

ISSN: 0809-8786

Department of Economics University of Oslo

Income Capability and Child Care

This series is published by the University of Oslo Department of Economics

In co-operation with The Frisch Centre for Economic Research

P. O.Box 1095 Blindern N-0317 OSLO Norway Telephone: + 47 22855127 Fax: + 47 22855035 Internet: http://www.oekonomi.uio.no/ e-mail: econdep@econ.uio.no

Gaustadalleén 21 N-0371 OSLO Norway Telephone: +47 22 95 88 20 Fax: +47 22 95 88 25 Internet: http://www.frisch.uio.no/ e-mail: frisch@frisch.uio.no

List of the last 10 Memoranda: No 13 Gunnar Bårdsen and Ragnar Nymoen

U.S. natural rate dynamics reconsidered. 28 pp. No 12 Enrico Sorisio and Steinar Strøm

Innovation and market dynamics in the EPO market. 32 pp. No 11 Hilde C. Bjørnland, Leif Brubakk and Anne Sofie Jore

Forecasting inflation with an uncertain output gap. 31 pp. No 10 Marina Della Giusta, Maria Laura Di Tommaso, Isilda Shima and Steinar

Strøm What money buys: clients of street sex workers in the US. 31 pp.

No 09 Cathrine Hagem and Hege Westskog Distributional constraints and efficiency in a tradable permit market. 27 pp.

No 08 Hilde Bojer Resources versus capabilities: a critical discussion. 16 pp.

No 07 Erling Barth, Alexander W. Cappelen and Tone Ognedal Fair Tax Evasion. 28 pp.

No 06 John K. Dagsvik Justifying Functional Forms in Models for Transitions between Discrete States, with Particular Reference to Employment-Unemployment Dynamics. 40 pp.

No 05 Hild-Marte Bjørnsen and Erik Biørn The Joint Labour Decisions of Farm Couples:A Censored Response Analysis of On-farm and Off-farm Work. 50 pp.

No 04 Morten Henningsen Moving between Welfare Payments. The Case of Sickness Insurance for the Unemployed. 35 pp.

A complete list of this memo-series is available in a PDF® format at:

http://www.oekonomi.uio.no/memo/

Income Capability and Child Care

Hilde BojerDepartment of Economics, University of Oslo,

POB 1095 Blindern, N-0317 Oslo, Norwayemail: hilde.bojer@econ.uio.no

JEL codes: D31, D63, I31Keywords: Income Distribution, Capability Approach, Cost of Children 1

15th May 2006

1Helpful comments from Halvor Mehlum and Charlotte Koren are gratefullyacknowledged.

Abstract

This paper has a twofold purpose. It argues firstly, that analysis of incomedistribution should be a part of the capability approach. Secondly, it arguesthat the capability approach is a good theoretical foundation for analysis ofincome distribution. The capability approach to human advantage has nothitherto been applied in mainstream analysis of income and wealth. Thepaper argues, on the one hand, that the capability approach needs to includea measure of access to economic goods; on the other hand that a capabilitydefinition of income avoids some problems inherent in the traditional, wel-farist approach. A modified full income is proposed as a measure of incomecapability. Modified full income is Beckerian full income corrected for interalia constraints on time. Time needed to care for children (and other depen-dants) is a constraint on available time. The capability approach will seek tocompute the cost of children (and other special needs) by identifying moraland legal constraints on time and other resources through open democraticdiscussion rather than econometric estimation.

1 Introduction

One purpose, and perhaps the main purpose, of analysing the distributionof income is to investigate whether the distribution is a just one. In order tomake sense in such a connection, the analysis should be related to a theoryof justice in distribution. In particular, the variable to be analysed ought tocorrespond to a magnitude whose distribution is relevant to justice. In em-pirical analysis of income distribution, it is generally taken for granted thatthe underlying variable of interest, the distribuendum of justice and of eco-nomic policy, is welfare, sometimes limited to ‘economic welfare’ or ‘materialwell-being’.1 In the political and moral philosophy of distributional (social)justice, this position is called welfarism.2 Amartya Sen has introduced theterm welfarism into contemporary philosophic debates. He defines welfarismas follows:

Welfarism in general and utilitarianism in particular see value,ultimately, only in individual utility, which is defined in terms ofsome mental characteristic, such as pleasure, happiness or desire.(Sen 1992: 6)

Sen is one of several contemporary philosophers who do not accept wel-farism as a proper basis for distributional and social justice.3 The first promi-nent critic of welfarism (or utilitarianism) was John Rawls in his famous ATheory of Justice (Second edition Rawls 1999). It is difficult to summarisehis detailed and profound criticism. One of his main points is the following:A liberal society is one that allows many different, even incommensurable,conceptions of the good in its citizens. Its distributional policy cannot there-fore be based on one single such conception, namely the pursuit of personalwelfare.4 Another notable contribution came from Ronald Dworkin (1981) in

1See e. g. Smeeding and Weinberg 2001.2The term welfarism was apparently coined by John Hicks in 1948 (reprinted 1981).3A concise overview of criticisms of welfarism is given in Bojer 2003 chapter 4. Note

that the criticism concerns welfarism in the normative context of distributional justice, andnot, or not necessarily, the assumption of utility maximisation when explaining behaviour.

4Modern utility theory includes the possibility of altruism (and envy) in the sense thatmy neighbour’s welfare may increase (or decrease) my own welfare. But the assumption ofthis kind of altruism does not break fundamentally with welfarism as a theory of behaviour.I still pursue my neighbour’s good because, and to the extent that, it makes me happy.It does not include the possibility that I do good from a sense of duty, and in spite ofconsiderations of my own welfare. Also, it is difficult to include altruism (and envy) in anormative theory of distributional justice with welfare as the distribuendum, since equalityof welfare would imply less to the altruist and more to the envious. The problem, in fact,is similar to that of expensive tastes discussed by Dworkin (1981).

1

his paper ‘What is equality? Part 1: Equality of Welfare’. One of Dworkin’sarguments against equality of welfare is the argument of expensive prefer-ences: some people are unhappy without private jets and Armani suits, oth-ers have a sunny disposition and are happy with very little. It does not seemfair, according to Dworkin, that those with modest demands should have tocontribute to the luxuries of those with expensive preferences. Like Rawls,Dworkin makes people responsible for their preferences.5 Now, in practice,welfarist analyses of income distribution generally assume equal preferences.Against this usage, Sen argues that there are objective differences in needswhich a just distribution should take into account. This argument forms partof the background for his capability approach.

The capability approach introduced by Amartya Sen and with MarthaNussbaum as a prominent advocate has aroused much interest and a largeliterature in recent years.6 Its ideas are applied in the UNDP’s Human De-velopment Index (HDI), and it has in general been influential in developmenteconomics. But the capability approach is so far more an inspiration thana developed theory. In particular, the capability approach has not, to myknowledge, been applied in mainstream analysis of economic inequality.

In the present paper I shall discuss how the capability approach canbe made a basis of defining and measuring income in analyses of economicinequality. I introduce the concept of income capability, and argue that theappropriate measure of income capability is a modified version of full income,where the modification includes constraints on time. This income conceptis in many respects an improvement on the standard definition of incomewithin the welfarist approach. It is in the definition and measurement of thecost of children that I advocate a break with the welfarist tradition.

Full income as the distribuendum of distributional justice also fits in withother contemporary theories of justice than the capability approach. In ATheory of Justice John Rawls advocated equality of ‘income and wealth’without elaborating further on the concept of income, but in a later paper(Rawls 1974) he indicates that full income fits in with his ideas. In Macro-justice (2005), Serge Christophe Kolm advocates a system of re-distributionindependent of hours worked, which implies that realised cash income cannotbe the basis of taxation.

I should, perhaps, make clear at the outset that it is not intended toreduce the whole capability approach to the one dimension of income capa-bility. All its adherents agree that there are many capabilities to be taken

5For an enlightening discussion of responsibility and welfare, se also Fleurbaey 1995.6The literature is too large to give more than a few selected references. I find that Sen

1999 gives the clearest exposition, while Sen 1992 includes his definition and criticism ofwelfarism. My own thinking also owes much to Nussbaum 1999.

2

into account in economic and distributional policy: the present argument isthat income capability should be one of them.

The paper is organised as follows: Section 2 discusses various standarddefinitions of income, in particular income as consumption and income asrights. Section 3 describes the capability approach while section 4 discussesthe relationship between capabilities and income. Section 5 introduces theconcept of income capability. Full income is reviewed in section 5.1. Insection 5.2 and section 6 I suggest how full income can be extended andmodified to become a suitable measure of income capability, and in particularhow to take the cost of children into account. Section 7 discusses the incomecapability of a household. In the last section, I discuss certain practicalaspects of the proposed theory.

2 Definitions of income

A generally accepted definition of income is that income is the upper limit ofwhat we can consume in a given period without diminishing wealth. In thewords of John Hicks:

The purpose of income calculations in practical affairs is togive people an indication of how much they can consume withoutimpoverishing themselves. Following out this idea, it would seemthat we ought to define a man’s income as the maximum valuethat he can consume during a week, and still be as well off at theend of the week as he was at the beginning. (Hicks 1961: 172)

This is the Hicks definition of income.Another much quoted definition of income is given by George Simons:

Personal income connotes, broadly, the exercise of control oversociety’s scarce resources. It has to do not with sensations, ser-vices or goods, but rather with rights which command prices (orto which prices may be imputed). (Simons 1938: 49)

.....Personal income may be defined as the algebraic sum of (1)

the market value of rights exercised in consumption and (2) thechange of value in the store of property rights between the begin-ning and end of the period in question. (Ibid.: 50)

The two definitions are usually taken to be identical. A close readingdoes, however, show an interesting difference between them. The Simons

3

definition stresses rights: rights to consumption, rights to property, and con-trol over society’s scarce resources. The Hicks definition concerns (possible)consumption only. Vilfredo Pareto’s discussion of income brings out the dif-ference:

Even the poorest man must be regarded as having sufficient in-come to keep him alive. It doesn’t matter whether this sum comesfrom the fruit of his work, or whether it comes to him from char-ity, or indeed whatever source, legal or illegal.7

According to Pareto the thief and the beggar both have an income. Butneither has rights in consumption, or (legal) control over scarce resources.They receive an income by the Hicks definition, but not by the Simons defi-nition.

The concept of income capability to be introduced below, is based on theSimons definition rather than the Hicksian one.

There is also the choice of period. The two standard periodisations are,of course, yearly income and lifetime income. In one sense, lifetime incomeis what we are looking for from the point of view of justice. John Rawls, forinstance, discusses the distribution of lifetime (economic) prospects. But thedegree to which annual income is an erroneous indicator of lifetime incomedepends on the degree to which people are able to choose circumstanceslike education and profession. And, as Atkinson (1983:42-43) points out,for people in precarious economic circumstances, shorter periods than a yearmay be of importance. I shall mostly have yearly income in mind, but alsoto some extent discuss the connection to lifetime income.

3 The capability approach

The capability approach is a theory for defining the quality of life, of humanadvantage. It has evolved in response to the shortcomings of welfarism as wellas to the narrowness of income, whether personal or national, as the dominanttarget of economic and development policy. Capabilities are independent ofthe preferences of the individual. In common with the Rawlsian concept ofprimary goods, the capability approach makes the individual responsible forher own preferences, and hence also for her own welfare. Both approachesalso include the possibility that the final good of human existence may not bepersonal welfare; indeed, it is up to each citizen to determine what is the finalgood, the goal, of her life. Therefore, the responsibility of the government

7Quoted from Brown 1976:75.

4

is not to secure the welfare of citizens; that is their own business. Theresponsibility of the government, and hence the aims of public policy, is tosecure for each individual citizen the capability to achieve her own goals,which may or may not be her own happiness or welfare.

The originator of the capability approach is Amartya Sen. He describescapabilities as capabilities to achieve valuable functionings. The qualifiervaluable here is important, since there are, conceivably, a great many desir-able functionings. Examples of valuable functionings are adequate nourish-ment and social participation. Sen uses the example of lack of nourishmentto illustrate the difference between a capability and the corresponding func-tioning. A starving person lacks the capability to nourish herself. A personwho fasts, does so voluntarily. She has the capability of nourishment, buthas chosen not to achieve the functioning.

Capabilities are individual. The unit for observing capabilities is alwaysthe individual person, not, say, the household. The philosophical basis forthe capability approach is ethical individualism. Ethical individualism doesnot imply that people have individualistic behaviour or preferences, still lessdoes it imply egoism as an ethical norm. It does imply that all persons areequally important from a moral point of view.8

Sen argues that the capabilities to be targeted by public policy in a givensociety should be decided on after, and by means of, open democratic deliber-ation, and refuses to lend his authority to any specific list of such capabilities.Martha Nussbaum has, on the other hand, published a list of minimum orthreshold capabilities that she considers applicable in all societies, but whichshe also considers as a proposal to be discussed.9 Some empirical surveys ofcapabilities have been carried out; these seem to have very similar ideas ofwhich capabilities are the important ones, as Ingrid Robeyns (2003) pointsout.

Income and wealth are not generally regarded as capabilities. Indeed,Robeyns writes: ‘...I deliberately exclude economic resources, as these donot constitute a capability. (Robeyns 2003: 73)’ By the term ‘economic re-sources’, Robeyns seems to mean either income or income and wealth, inaccordance with a fairly widespread, but in my view unfortunate, usage.10

But it is hard to see how any analysis of distribution, or any distributionalpolicy, can avoid treating access to economic goods in one form or another.One reason is the fact that taxes and transfers are important instrumentsof distributional policy. They will not lose importance in a policy that tar-

8On ethical individualism, see e.g. Dworkin 2000:446-4529See Nussbaum 1999 and 2000.

10She is more precise in Robeyns 2005.

5

gets capabilities, since capabilities are not themselves transferable from oneindividual to another; nor can they in most cases be directly controlled ormanipulated by the government. Sen repeatedly stresses that income is in-strumental to capabilities. Precisely, and that is why we need to investigatethe role income should play in the capability approach.

Sen clearly regards income as a means of acquiring commodities, whichare inputs to be transformed into capabilities. On the other hand Nussbaum(1999:42) has the right to own property and to engage in paid work on herlist of threshold capabilities. This right seems to correspond to a capabilityof earning an income. The relationship between capabilities and income istherefore far from straightforward.

4 Capabilities and income

I now turn to Amartya Sen’s discussion of the relationship between incomeand capabilities in Development as Freedom, chapters 3 and 4. Here, hewrites: ‘...income - properly defined - has an enormous influence on whatwe can or cannot do. (1999: 72)’ But he also identifies ‘at least five sourcesof variation between our real incomes and the advantages - the well-beingand freedom - we get out of them. (1999: 70)’ These are: (1) distributionwithin the family, (2) differences in relational perspectives, (3) environmentaldiversities, (4) variations in social climate and (5) personal heterogeneities.11

He does not, however, tell us on which definition of income he bases the list,even though he stresses that income must be ‘properly defined’. The contextmakes it sufficiently clear that what he has in mind is income as a measure ofpossible consumption, that is, the Hicksian definition of income. Moreover,he also seems to have in mind cash income, which measures, more or less,access to market commodities.

Sen’s criticism of income can be grouped into two kinds. One concernsincome as a measure of possible consumption, the second concerns consump-tion as a measure of capabilities. Distribution within the family and en-vironmental diversities I would class as concerning income as a measure ofconsumption. The physical environment is an economic good, and a publicone. The argument of environmental diversities shows that access to indi-vidual commodities is not the only indicator of possible consumption, andshould be supplemented by other indicators. But it does not indicate thatincome is a faulty measure for consumption of individual economic goods.(market commodities).

11The numbering is mine, and introduced for ease of reference.

6

The argument of distribution within the family is of central importance,and I shall return to it below.

The remaining items concern the relationship between consumption andcapabilities. By the second item, ‘differences in relational perspectives’, Senmeans that many valuable capabilities depend on our affluence relative tothat of the rest of society. The point is well known from the literature onrelative poverty, but does not only apply to people below the poverty line.It is one of the most important insights of the capability approach, withprofound consequences for policies of development and growth. But it doesnot make superfluous the measuring of income; on the contrary, for relativeincomes to be known, absolute incomes must first be measured.

‘Variations in social climate’ I interpret as concerning rights and liberties,not access to economic goods.

The last item on the list is (5) ‘personal heterogeneities’. The dissim-ilarities (heterogeneities) refer to needs, not (necessarily) preferences; thedifference being that needs are conceived of as objective and observable. Astandard example is special dietary needs due to for instance pregnancy. Ishall argue that a reasonable number of such observable special needs can bedealt with by suitable numerical corrections of income, to be discussed belowin section 5.2.

To Sen’s list of ‘sources of variation between our real incomes and theadvantages - the well-being and freedom - we get out of them’, I would addat least two other items: leisure and publicly provided individual goods.These are both items of individual consumption.

Leisure (item 6): The deprivation of a man who has to work 12 hours aday to keep body and soul together is inadequately measured by the smallnessof his cash income. Also, many women have the responsibility for childrenand housework in addition to wage labour. So, a given cash income may beassociated with very different levels of leisure. In particular, the fact that aperson is without any cash income at all may be due to leisure, enforced ornot, but may also be due to unpaid work and very little leisure.

Publicly provided individual goods (item 7): Some goods are not tech-nically public in the sense of being non-rivalling, but they are neverthelessprovided wholly or in part by the government in many countries. Importantexamples are social security, health care and education. Such goods havecomputable values that should be added to income to the extent that theyare paid for by the employer or the government. There is obviously a greatand important difference between a cash income that has to cover social andhealth insurance and one that does not.

Sen also does not mention home production, which many authors (e.gJenkins et al. 1998) now include in their measure of income. I shall discuss

7

the inclusion of home production below in section 7.

5 Income capability

it seems to me that several of Sen’s criticisms of income as well as my owntwo additions to the list, can be dealt with either by suitable corrections ofstandard cash income (5 and 7) or by observing suitable indicators of thestandard of living and human rights in addition to income (2, 3 and 4). Thetwo items that demand a serious revision of the concept of income, seem tome to be those of leisure and of distribution within the family. In addition,there is the aspect of income that has to do with power and rights, whichSen does not really touch upon in his discussion, but which is one reasonwhy Nussbaum includes the right to own property and earn an income inher list of basic capabilities. Sen also writes about unemployment as beingdeprivation of a capability, not of income only.

I shall consider first leisure, then distribution within the family and finallyincome as power before concluding by defining income capability.

In welfarist analysis, the difference between work and leisure is thatleisure yields utility, while work does not. Now, not everyone enjoys leisure,while some people enjoy their work. Unemployment, for instance, is enforced‘leisure’, a loss of capability, as Sen points out. From the capability pointof view, leisure must be defined not as time yielding utility but as free time,that is, time when you are free to choose what to do. Also, it is leisure as anoption, not a realised fact, that is important to the capability approach.

When studying consumption, distribution within the family is of courseof great importance. It is, however, by now well known that this distributioncannot be assumed to equalise the consumption of family members. Thenorms governing within-family distribution vary not only from country tocountry, but also within the same country. Amartya Sen was one of thefirst economists to discuss within-family (or -household) distribution. In Sen1984, he describes how girls from two neighbouring villages in India werevery different with respect to health and physical development because ofdifferences in the internal allocation of food in the households. The girls inone village received a smaller part of the household’s food. The necessity ofmonitoring the standard of living of parents and children, girls and boys, menand women, separately is receiving steadily more attention in developmenteconomics. So is the value of targeting aid and development efforts directtowards the education and income making opportunities of individual women,not their households (husbands). Helping women out of poverty is also the

8

surest way of helping children out of poverty.12 The knowledge that growth inproduction and rising standards of living is unequally distributed both withina country and within each separate household forms part of the backgroundfor the capability approach and the construction of the Human DevelopmentIndex.

In advanced, Western countries, there is reason to believe that distribu-tion and decision making in general within the household is influenced by thebargaining power of the individual members, which in turn is determined bytheir individual economic opportunities outside the household. (Lundbergand Pollak 1993).

Inequality of distribution within households is a reason for monitoringthe standard of living of household members separately (as far as possible),as well as providing for individuals separately in distributional policies. Foradults, there is a strong case for measuring individual incomes instead of, orat the very least in addition to, household income.

The bargaining model of household decisions also reveals why income isa source of power. But the income that gives power is the income that stemsfrom legally and socially enforceable rights such as : the right to paid labourat a certain wage, legal rights to social security, child benefit or alimony asthe case may be. Handouts from charity are not rights.

Sen has often written that the important thing is not the income a personenjoys, but what she is able to do and to be. ‘Capability is thus a kind offreedom: the substantive freedom to achieve alternative functioning combi-nations (or, less formally put, the freedom to achieve various lifestyles). (Sen1999: 75)’ Within the capability approach it is the potential income, theincome a person might have chosen had she so wished, which is of interest.Consumption of commodities and hence income as a measure of consumption,is a functioning, not a capability.

The capability approach to income should therefore measure the capabil-ity to acquire an income, not the income itself. This capability is determinedby legally and socially enforceable rights to non-labour income, the size ofthe obtainable wage, and freedom from time constraints. Considerations ofpower and the possible inequalities in within-household distribution lead todefining income capability as an individual capability.

The closest approximation I can find to income capability understood asthe option of choosing between cash income and leisure is individual full in-come, determined by available time, the wage rate and by non-labour income;and suitably corrected for the sources of ‘variation’ discussed above: specialneeds and government paid individual goods like social and health insurance.

12See for instance Todaro and Smith 2005 chapter 5.

9

The non-labour income part consists of capital income and government orprivate transfers guaranteed by law.

Obviously, this notion of income capability is only applicable to adults.Children’s capabilities are another matter; I shall discuss some aspects oftheir economic position in section 7 below.

If income capability measures options, possible choices, it is importantto have in mind that the options any given year will to some extent bedetermined by past choices, at least for some persons in some societies. Thislast reservation is made because the majority of the earth’s inhabitants do notreally have many options to choose between. They cannot choose betweenworking and not working, marrying or not marrying, never mind choice ofeducation and profession. This is particularly true for women.

I shall discuss the income capability corresponding to annual income.Then, the constraints determining capability will inevitably at least partlybe due to previous choices made. But the same is true of standard annualcash income.

5.1 Full income

The concept of full income is not new; it was introduced by Gary Becker(1965). Full income as a theoretical concept is widely applied in labourmarket economics and fertility analysis, but has not to my knowledge beenempirically measured and applied in analysis of economic distribution. Thedifference between full income and standard income is that full income in-cludes leisure.

In symbols:FI = wT + MI (1)

where w is the wage rate, MI is non-labour income and T is availabletime, say 16 hours a day.

Let two persons, A and B, have equal non-labour incomes (MI = 0), butdifferent wage rates, so that WA > WB. The relationship between their cashincomes (I) may well be different: IA < IB because they have different hoursof paid work.

Gary Becker, and after him economic theory, treats these different out-comes as the results of free choice between paid work and leisure, determinedby the preferences of a utility maximising consumer. In the capability ap-proach, the motives for the choices made are not important. What is impor-tant, is the capability of choosing, in other words, we must seek to identifythe constraints on choice.

10

And there are other constraints on choice than those represented by equa-tion (1): constraints on time and constraints on consumption. Working hoursmay be fixed by the employers, so that the only choice is between havingpaid work full time or not at all. There are social, more or less informal butnonetheless real and strict, constraints on how both women and men shouldspend their time. For women, particularly mothers, having paid work maybe frowned upon. Conversely, employers may expect male workers to workovertime regardless of obligations to the family.

The assumption of a constant wage rate is also unrealistic. It does not,for instance, take into account extra pay for overtime, nor varying tax rates.

Full income is determined partly by personal endowment (human capital),measured by the wage rate, partly by social, economic and legal features ofthe surrounding society. These features of society determine to which extentequation 1 is the only constraint on choice. In particular, a great numberof persons, children apart, in any society have no choice at all between paidwork and not paid work. For disabled and retired persons it is natural toset the wage rate equal to zero: they are dependent on pensions, or, in somesocieties, charity. I would suggest the same procedure for the unemployed.For persons that are unemployed and able-bodied and of sound mind, itmight be reasonable to add a certain sum, or assume a small wage rate, totake the possibility of home production into account. Married women incultures where they are denied the possibility of paid work would also havea wage rate equal to zero, and hence no income capability.

I shall discuss three types of constraints on time and consumption: spe-cial needs, consumption costs of children and socially necessary unpaid work(time cost of children). Some, or all, of these constraints, may be due toprevious choices. The case of children is in particular the subject of muchdiscussion. But the wage rate may be influenced by decisions on educationand career choice, and special needs may be due to risks deliberately in-curred. These problems are no different in the case of income capability thanwhen measuring standard cash income. It is for example regular practice tocorrect household income per equivalent adult for economies of scale in thehousehold: but marriage and cohabitation are as much voluntary decisionsas having children is.

5.2 Special needs

The standard example of different needs used by both Sen and Nussbaum isdifferences in nutritional needs. These are, of course, of great importance topeople living on or near subsistence level. But there are many other casesof different needs which are of importance at all levels of wealth: differences

11

due to chronic illness and/or physical and mental handicaps. This is wherepeople need different quantities and qualities of commodities in order toobtain basic capabilities: some need glasses in order to see, others needcrutches or prosthetics in order to be able to move.

These needs can be measured in money, namely the cost of medicines andphysical equipment needed. The monetary values of commodities needed aregenerally already known, more or less approximately, since in many countries,the supplementary costs are covered by governmental or private insurance,at fixed rates. There is a limited, though large, number of different cases.In each case, the necessary sum must be deducted from full income in orderto compute the individual’s real options. The sum to be deducted is theperson’s outlay net of public subsidies where these apply.

To be sure, it may in some cases be impossible to achieve normal capabil-ity, and sometimes the achievement of a basic capability may be prohibitivelyexpensive. This is one of the unsolved problems of the capability approach.

Medication and other medical aids may enable the person to undertakepaid work, and are therefore both a result of, and an input into, incomecapability.

6 The cost of children

6.1 Consumption costs

Children cost time and money: time that the parents have to spend on theircare, and money spent on the children’s consumption. In economic analy-sis, children are mostly considered either as individual goods acquired bythe parents for their own pleasure (Becker 1991), or as public goods (Folbre1994). In analysis of justice in distribution, they are neither, but humanbeings with rights of their own. Hence, the cost to parents stem from thechildren’s rights to care and consumption, and the parents’ ensuing obliga-tions to provide these. Acquiring children may in modern society be a freechoice. Spending a minimum of care and money on the children, once theyare there, is not. Parents have obligations towards their children; moralobligations everywhere, in most societies also legal obligations. Moreover,these obligations are independent of the pleasure parents may or may nottake in being parents. There is therefore no reason to distinguish betweenthe conditional and unconditional cost of children; these concepts arise outof the welfarist approach.13

13The unconditional cost of children is defined as the welfare loss to parents due tochildren’s consumption less the welfare gain obtained from having children.

12

There is a considerable literature on the problem of determining the con-sumption cost of children. The results vary widely, and there is no consensuson what the right answer is. I would suggest that the reason for the lackof consensus is that there is no answer that is, objectively, the right one. Ithink we should get further by owning outright that determining the cost ofchildren is an evaluative exercise. It is the answer to a normative question,namely: how much ought society allow for children’s consumption.

The main body of the existing empirical literature is devoted to comput-ing the sums that parents actually, on average, spend on their children. Butactual expenditure is an outcome of choice. Children cost the amount of timeand money that parents are able and willing to spend on them. This averageexpenditure will, of course, vary from society to society, depending both onthe society’s wealth and on accepted social norms. Sen’s observation in Sen1984, referred to above, showed that parents in one village gave more foodto their sons than to their daughters. There is reason to believe that such anunequal distribution is not uncommon in India. Could observed expenditureof this allocation be morally acceptable as basis for computing the cost ofsons and daughters?

The capability approach to evaluating the cost of children would be toacknowledge the normative content of the exercise. It would seek to estimatethe constraints on parents’ choice given the presence of children. Since we arelooking for moral and legal constraints, the costs should not necessarily beestimated from observations of parents’ actual spending. Sen claims that anadvantage of the capability approach is that it is open for democratic discus-sion. The capability cost of children might well be decided on in this spirit.I tend to be of the opinion that the cost in terms of commodities should beindependent of the parents’ income, and should represent societal consensus(or majority opinion) on what the minimum legal and moral obligations ofparents are towards their children.14

Both empirical investigations of actual expenditure and experts’ estima-tions of necessary cost would be valuable inputs into normative deliberations,but it should be made clear that they are inputs only, and that there is no‘right’ amount to be scientifically established.

6.2 Time cost of children and unpaid work

The time cost of children will depend on the wages of the parents. However,the time itself, the number of hours per day or days per year, necessary for the

14It is interesting that in the article that spawned the whole literature on conditionaland unconditional cost, Pollak and Wales (1979) recommend the estimation of children’sminimum requirements.

13

care of children, is independent of wages or income, and should be determinedby methods similar to those recommended for necessary consumption. It willvary with the age of the child, from 24 hours a day for the new born infant,and decreasing with the age of the child, but never becoming zero as long asthe child is a minor.

Unpaid work is one constraint on available time. Now, unpaid work is, ofcourse, sometimes chosen voluntarily, and may be difficult to distinguish fromleisure pursuits (hobbies). The criterion used in empirical work to distinguishunpaid work from leisure is the third-person criterion: an activity is unpaidwork if it is technically possible to have someone else do it for you, paid orunpaid. Baking bread is unpaid work, a work-out is leisure. Another possiblecriterion, more difficult to observe, is the possibility of choice: leisure is freetime, when you have a free choice of activities. Time spent on unpaid workis, by this criterion, time when you have obligations toward others, timewhen you have no choice but to work. The two criteria do not, in general,coincide. The baking of bread may be undertaken as a hobby, but may alsobe a necessary duty.

Caring for children and infirm or sick dependants is, however, work ac-cording to both criteria. It has to be done by somebody, whether the personcared for is able to pay for the care or not. Parents must either care for theirchildren themselves or pay someone else to do it for them. Indeed, leisure,free time, is scarce for parents in modern society.15 Children’s need for timemeasured in hours is absolute, and does not depend on the parents’ income.The money value of the time will, however, vary with the real wage of theparents. Time needed for child care consists of two parts: necessary timespent by the parents and time hired from a third person (day care centre,nanny). Both must be deducted from time available when estimating incomecapability.

Let me assume for the moment that there is just one parent. The priceof the parent’s time will be his or her own wage rate. The third-person timewill be priced at the market wage for child carers. This wage may, of course,be higher than the wage rate of the parent.

Now, let p be the price per hour of outside child care, and w the parent’swage rate, T0 is necessary parent time, T1 is third person time. The care costof a child will now depend on whether w > p or w < p. If w > p, the parentis able to earn her own income and use part of it to pay for child care. Butif w < p, she will not be able to afford outside child care.

CT = wT0 + pT1 if p < w (2)

15See e.g. Vaage 2002: 184

14

CT = wT0 + wT1 if p > w (3)

Equations (2) and (3) express the care cost in money. It can also berepresented in terms of time, the time a parent must work to cover herchild’s need for care:

TC = T0 +p

wT1 if p < w (4)

TC = T0 + T1 if p > w (5)

In the above equations, I have assumed that the parent will care for thechild herself if p > w. Above a certain age, this is not necessarily the bestsolution for the child. Again, the best division of care time between privateand professional should be decided by open public discussion. Also, I havenot taken into account the fact that it is often technically possible to combinechild care at home with other activities. This fact might be represented bysubtracting a sum from home care cost.

The total cost of children is the sum of consumption cost and care cost,which can again be computed in time or money.

The income capability of adults should be measured individually. Then,the cost of children (and other dependants) must be individually assigned.There is no simple answer to how this should be done when a child liveswith both parents. Immediately after the birth and during breastfeeding,the care time must be assigned to the mother and deducted from her fullincome. Later, there is no biological reason why the time cost should notbe equally divided between the parents. In many countries, social normswould perhaps support the equal division. In other countries, social normsunequivocally call for the mother to spend time caring for children. Even incountries where many people feel that father and mother have equal moralobligations to spend time with the children, the time burden in practicefalls on the mother. I suggest parents be interviewed about how they dividethe time spent caring for children, and that the time cost be individuallyassigned according the answers received. The consumption cost it seemsmost reasonable to divide equally.

7 Children and the household

Income capability as I have defined it, does not make sense for children. Chil-dren are not expected, in most modern societies, to earn their own income.Still, economic provision for children is an important part of distributional

15

policies, again in most societies. And children are, of course, dependent onthe parents’ ability (and willingness) to provide for them. When investigatingthe economic conditions of children, it is therefore necessary to investigatethe economic conditions of the parents, providers or provider, as the case maybe. We can only get to know the economic position of children by studyinghow capable the parents are of supporting their children economically. Hencethe need for a concept of household or family income.

There is a case, however, for modifying the concept of household incomealong the same lines as I have advocated for individual income capability.The standard measure is, of course, disposable household cash income perequivalent adult. Equivalent adult scales are normally based on consump-tion cost only (and economies of scale in consumption), and do not includethe time cost of children, (where there are also possible economies of scale).Moreover, a given cash income will give different consumption possibilitiesdepending on time available for household production. There is a consider-able difference between 50 000 Euros earned by two partners in full time jobsand the same income earned by a husband with a full time housewife. In thelast case, there will be her home production to add to the consumption ofthe couple.

But I think full income will be a better measure of their consumptionpossibilities than so-called extended income which includes household pro-duction. Household production may be at the cost of leisure. Two householdswith equal extended incomes may well have different full incomes; the differ-ence is due to different hours of leisure. For these reasons, when it comes tocouples, the sum of their individual full incomes will usually give a truer pic-ture of their combined consumption possibilities than does either the sum oftheir cash incomes or their total extended income. In other words, I suggestwe measure the sum of their income capabilities as previously defined, whereinter alia the full cost of the children is deducted. This magnitude we mightcall the income capability of the household. This capability is the differencebetween the parents’ total income capability and the socially acceptable con-sumption and time cost of their children. The difference may well in somecases be negative, particularly for single parents.

I should, however like to make a reservation for the case of societies wheremarried women for legal or cultural reasons have no economic options outsidethe family. It would not be reasonable to set equal to zero their contributionto household consumption.

Income capability is particularly suited to express the real difference ineconomic circumstances between single providers and couples with children.The couple will have twice as much time as the single provider has to dividebetween caring for children and earning an income.

16

It is important to make clear that what I called the income capability ofthe household measures the capability, that is, the possibility of the parentsto consume commodities and leisure after having provided for their childrenaccording to accepted standards. Like the conventional measure of householdincome per equivalent adult, it says nothing about what parents actually do.The life quality of children must be monitored separately.

8 Some practical considerations

All the data needed to compute income capability are of the kind alreadycollected though household panels and in labour force surveys. In particular,the questions needed are neither more nor less intrusive. More attentionwould be needed to constraints on time, like the questions now asked in manylabour force surveys about preferences for shorter or longer hours. Questionswould also be needed about the wage rate in present or, for the unemployedand for homeworking housewives, past occupations. Couples with children(or other dependants) should be interviewed about how they share time spentin care work.

Finally, some words on the feasibility of deciding on the rights of childrenin democratic deliberation instead of by expert calculations. The so-calledcosts of children are, in fact, administratively and/or politically decided onin many cases in connection with tax schedules, child benefits, welfare con-tributions and alimony. Arguments that these are too small, or too large,are, I feel, better based on considerations of rights and social justice thanon scientific, expert judgements. Also, acceptable standards will vary fromsociety to society, depending both on the general affluence and on the socialand political mores of that society.

Democratic deliberation on standards is an important part of the capa-bility approach.16 Amartya Sen writes:

Furthermore, the need to discuss the valuation of diverse ca-pabilities in terms of public priorities is, I have argued, an as-set, forcing us to make clear what the value judgements are ina field where value judgements cannot be - and should not be- avoided. Indeed, public participation in these valuational de-bates - in explicit or implicit form - is a crucial part of the exerciseof democracy and responsible social choice. In matters of pub-lic judgement, there is no real escape from the evaluative need

16An excellent discussion of this aspect of the capability approach is found in severalarticles in Maitreyee 2006.

17

for public discussion. The work of public valuation cannot bereplaced by some cunningly clever assumption. (Sen 1999: 110.)

9 References

Atkinson, Anthony B (1984), The Economics of Inequality, ClarendonPress, Oxford.

Becker, Gary (1965), ‘A Theory of the Allocation of Time’, EconomicJournal, 75: 493–517.

Becker, Gary (1991), A Treatise on the Family, Harvard University Press.

Bojer, Hilde (2003), Distributional Justice. Theory and Measurement, Rout-ledge.

Brown, J. A. C. (1976), ‘The Mathematical and Statistical Theory of In-come Distribution’ in A. B. Atkinson (ed.) The personal Distributionof Income, Allen and Unwin, London:72-97.

Dworkin, Ronald (1981, ‘What is Equality Part 1: Equality of welfare’,Philosophy and Public Affairs, 10: 185–246.

Dworkin, Ronald (2000), Sovereign Virtue, Harvard University Press.

Fleurbaey, Marc (1995), ‘Equal Opportunity or Equal Social Outcome?’,Economics and Philosophy 11: 25-55.

Folbre, Nancy (1994), ‘Children as Public Goods’, American EconomicReview, 84(2):86-90

Hicks, John (1961), Value and Capital, The Clarendon Press, Oxford.

Hicks, John (1981), Wealth and Welfare, Basil Blackwell, Oxford.

Jenkins, Stephen, P. M. Kapteyn and B. M. S. Van Praag (Eds) (1998),The Distribution of welfare and Household Production: InternationalPerspectives, Cambridge University press

Kolm, S. C. (2005), Macrojustice. The Political Economy of Fairness, Cam-bridge University Press.

Lundberg, Shelley and Robert A. Pollak (1996), ‘Bargaining and Dis-tribution in marriage’, Journal of Economic Perspectives 10: 139-158.

18

Maitreyee 2006, Briefing of the Human and Development Association num-ber 4, February, available on http://www.hd-ca.org.

Nussbaum, Martha C. (1999), Sex and Social Justice, Oxford Univer-sity Press.

Nussbaum, Martha C. (2000), Women and Human Development. TheCapabilities Approach, Cambridge University Press.

Pollak, R. A. and T. J. Wales (1979), ‘Welfare comparisons and equiv-alent scales’, American Economic Review, 69: 216–221.

Rawls, John (1974) , ‘Reply to Alexander and Musgrave’, Quarterly Jour-nal of Economics, 88:633-655.

Rawls, John (1999), A Theory of Justice. Revised Edition, Harvard Uni-versity press

Robeyns, Ingrid (2003), ‘Sen’s’s Capability Approach and Gender In-equality: Selecting Relevant Capabilities’, Feminist Economics 9: 61 –92.

Robeyns, Ingrid (2005), ‘Assessing Global Poverty and Inequality: In-come, Resources and Capabilities’, Metaphilosophy 36(1-1):30-49

Sen, Amartya K. (1984), ‘Economics and the Family’, in Amartya K.Sen, Resources, Values and Development, Basil Blackwell, Oxford: 369-388.

Sen, Amartya K. (1992), Inequality Re-examined, Clarendon press, Ox-ford.

Sen, Amartya K. (1999), Development as Freedom, Oxford University Press.

Simons, H (1938), Personal Income Taxation, University of Chicago press,Chicago

Smeeding, T. M, and D. H. Weinberg (2001), ‘Toward a uniform de-finition of household income’, Review of Income and Wealth, 47: 1–24.

Todaro, Michael P. and Stephen C. Smith (2005), Economic Develop-ment, Pearson-Addison Weasly.

Vaage, Odd Frank (2002), Til alle døgnets tider: Tidsbruk 1971-2000(At all hours of the day. Time use 1971-2000), Statistiske Analyser52, Statistics Norway.

19